ARCHIVED FORUM -- March 2012 to February 2022READ ONLY FORUM
This is the second Archived Forum which was active between 1st March 2012 and 23rd February 2022
Well, after having my M100's phaselinks repaired my Beomaster 8000 refuses to startup properly. It does display the decimal dot, and does click it's relays when pressing "0", and responds properly to the volume control and other commands, but within a minute it freezes with just "0" in the display.
Odd: when I unplug the machine and try again immediately no button responds, but when I leave it off for a few minutes I can repeat the cycle.
Pressing "0" and the Monitor button simultaneously to start the test program correctly displays "TP" (Test Passed), if I press those buttons quickly enough. If I'm too late, after half a minute no button responds.
I unplugged all cables to other circuits so I only have the computer and it's display, and I still get this behaviour, so my assumption is the problem is located in the computer board.
Found http://beolover.blogspot.nl/2016/10/beomaster-8000-rapid-relay-switching.html where sort of similar behaviour is described. Apparently one remedy is to replace the 2MHz crystals, and re-solder VIA's. My board seems to be an older one than most described (mine is 1980, judging the chips date stamps), there are only very few of those on my board. But, the more extra components soldered on. Gosh, such an ugly board:
Re-soldering the VIA's with a bit of wire did not help, now awaiting delivery of the crystals, as my local electronics supplier did not have them in stock (grmbll).
If all else fails what I will do is simply replace the computer board with an Arduino. I see the protocols for the display are described well, and also controlling the amplifier can't be difficult. The volume and radio frequency are just rotary encoders, and those are not difficult to read using Arduino.
Arduino also has 1024 bytes of EEROM storage, so I can store frequencies there too.
Stay tuned...
rick
Interesting though I think I won't be as simple as it may appear...
Many CPU problems are down to bad filtering of the 5V supply.Check (or replace) the capacitor on the right side cooling, where the voltage regulators are.
Martin
Thanks! Tried that, but no. I do notice the 5V is 4.82V in reality, so 3.6% less than 5V.
Perhaps I should replace the 5V regulator, it's just a 7805..
Replaced the 7805, voltage now is 4.98V, a bit closer to 5V, but still: nope.
Really odd thing there is something with a memory in this process. If I unplug and replug in, say 5 seconds, the Beomaster computer still is unresponsive. If I retry a minute later the processor works fine, for about 30 seconds, before hanging again.
This is consistent behaviour. So there is something with a memory that gives up after ca 30 seconds. And it's still messed up if you replug and retry too quickly.
Replaced C85, the 22uF capacitor on the processor board, too, still no.
Dillen:Interesting though I think I won't be as simple as it may appear...
Reading the Tech.Produktinfo now (German), and it does not seem that horrible. It'll be quite some work, but it'll be fun too!
- the volume control is simply a 60 steps (6-bit) pattern, stored in the computer, controlling a volume ic. That pattern is in/decremented with the rotary encoder volume.
- the tuning is not as "digital" as you think it is, it is basically an analog tuning voltage generated by the microcomputer, by integrating an PWM signal generated by the computer into a voltage, then reading back the actual tuning frequency by measuring that, and adjusting while no match. Either controlled with the tuning rotary encoder, or by the computer. When the computer sets the tuning frequency it also sets a "fast" signal on the integrator circuit, so the tuning is achieved much faster than by spinning the wheel. It's all analog though :-) Generating PWM signals is something Arduino is very very good at. That old 6500 in the Beomaster is getting warm, I think I now understand why.
- FM frequencies are stored as a (serial) bit pattern in a TMS3529 non-volatile ram ic, controlled by the computer. (It uses a 33V programming voltage, switched on by a transistor, controlled by the computer.) Odd, I haven't found a single datasheet yet, though it's made by Texas Instruments.
- the display has basically 3 multiplexed normal 7-segment decoders, and a few extras. A fun extra is that they needed to display "PH" for PHONO, but the 7-segment decoder cannot display a "H". So what they did was display a "4", and light one extra segment with a dedicated transistor.. haha, great stuff!
- a bit special is that the various devices (beomaster, beogram, beocord) are linked, and they can control each other by sending serial bit patterns.That is what the extra pins in the DIN connector are for!
- the keyboard and the switches on the other panel (monitor, filter, manual/auto tune etc) are read by scanning/multiplexing with the computer (3 columns, the same as the display 7-segment multiplex signal, 7 rows).
Way back when I determined that my BM5000 had a dead uC chip, I almost started a similar project. I was going to use a Teensy board though. Then I found an original replacement processor board and used that - it would have been impossible to implement 2-way remote control without a working system as a reference anyway. At least you don't have that problem, Good luck - I'm following with interest!
--mika
- Replacing the crystals did not fix it. Well, €1,50 down the drain. At least the processors still work.
- made sure the remote control signals could not reach the processor, in case that circuit is messed up, but no.
- unsocketed the TMS3529, that also has a direct connection to the processor, but, again no.
- grounded the input of the RESET schmitt trigger, so RESET=1: nope.
What now? Replace is the HEF40106 6-schmitt trigger? That is about the only thing left supplying the processor with inputs... (The tuner is disconnected, the rest of the processor board controls the display.)
(eyeing the Arduino Mega on the shelf here)
Look at Tim's solution!
http://www.vintage-radio.net/forum/showthread.php?t=110262
Jacques
Wow, he has replaced the individual 6500 processors!
I intend to replace the entire processor board with an Arduino, unless I get it fixed after all. That board is horrible anyway.
Arduino Mega has 54 digital pins, or I can use a regular Arduino with an I/O expander, e.g. with an I2C bus. Or, I'll use two Arduino Mega's.. (Arduino boards are so cheap these days!)
@Dillen: Saw a post regarding the prescaler in the archived forum: http://archivedarchivedforum2.beoworld.org/forums/t/28115.aspx asking wether the u264b and u664b prescalers are equivalent, Well, my Beomaster 8000 has a u664b, so the answer is yes.
The prescaler is something I need to keep, or copy, to measure the tuning frequency. So perhaps I will leave the processor board in place when switching to Arduino, although stripped.
Um, my schematic has just one ic8 HEF4013 dual FF for further dividing by :4 the :64 FM frequency generated by the 664, but on the actual board there's also a floating board with another HEF4013.. Does that further divide by :4 ? It does seem so, as the output of the floating board goes to pin 21 of the slave processor, which is the frequency counter pin, with the trace going directly to ic8 severed. So, not divided by 256, but by 1024? It would make the work for the slave processor easier.
I'll try replacing the two HEF4013's and the HEF40160 next.
rickjansen: @Dillen: Saw a post regarding the prescaler in the archived forum: http://archivedarchivedforum2.beoworld.org/forums/t/28115.aspx asking wether the u264b and u664b prescalers are equivalent, Well, my Beomaster 8000 has a u664b, so the answer is yes. rick
So that you have one that works confirms that they are identical?What about the handful or more U664Bs that I've had that didn't work, where U264Bs did - doesn't that confirm that it's something that cannot be guaranteed?
And yes, during production it turned out, that some processors simply couldn't keep count at the higher frequency, so to keep it safe anadditional divider was introduced and the CPUs internal code changed to suit.
Did you test your 664's later? Are they really faulty? Could very well be, these prescalers seem to break quite often, if I can believe Google. Can't find a datasheet for 664 and 264 yet... I'm just reporting my machine has a 664, and it worked fine.
I haven't tried the U664Bs in other circuits and I haven't looked much into the technical details butthe U664Bs are probably not faulty, it's just a different IC, functioning as intended but not quiteright for the given task for reason(s) unknown to me.The problem is the same - only far more common - in Beomaster 6000, where the IC sits in awarmer area, right up against a voltage regulator.There's not a lot of room in there, but I like to bend the regulator away from the prescaler IC as muchas possible for the same reason.I have replaced three or four prescaler ICs in Beomaster 8000s and lost count of the onesin Beomaster 6000s - but literally dozens, perhaps a hundred.I have seen one single U664B that seemed to be factory mounted (and replaced it because it had stoppedworking), it was in a Beomaster 6000 on my bench recently. I have also heard about a few other repairers that had "luck"with U664Bs but far more had problems like I.The original TFK (Telefunken) U264Bs works every time in both Beomasters and I've supplied ICs toseveral DIY'ers that had problems finding a replacement that worked.(Other brands U264Bs cannot be guaranteed to work either).Martin
Dillen: U664Bs vs U264Bs
U664Bs vs U264Bs
Interesting, and thanks for the explanation!
Regarding heat: the power board of my Beomaster 8000 looks quite "fried", the area of the board between the two power resistors is quite darkened (is that normal?):
I resoldered R56 and R55 to hover about a millimeter above the board; they were touching the board before!
It must mean the board above this (the processor board) is at an elevated temperature too, although the metal shielding there may have helped to prevent damage.
Your board looks slightly more cooked than average.But it's not uncommon to see it like this.
Datasheet for u264b: http://novatech-instr.com/Fun/u264b.pdf
For u664b: http://www.datasheet.hk/view_download.php?id=1435101&file=0200%5Cu664b_4564098.pdf
It seems max input voltage for u264b is 500mV, for u664b it's 0..Vs. Some circuits using the chips are drawn with Shottky clamping diodes on the inputs, which seems a sensible precaution. I don't see those in the Beomaster 8000, or I've overlooked them so far.
Hi,
For the record, I have a 664 in my BM 6000.
Well, today I replaced the HEF40106 and the two CD4013's. The HEF40106 cost me a bicycle ride twice, because the shop was out of HEF40106's, and I only later realized that the 74HC(T)14 is a direct replacement here (5V Vss). They did have the 74HC14.
But, still: nope.
I unsocketed the 6500 master, as there are solder points under that, to see if there are any VIAs left to resolder there. But, the socket is not open, so I cannot get at the board there! Resocketed the 6500, and the board still "works" as before, so I have not damaged anything as far as I can see.
I'm a bit out of ideas at the moment. I'll do some more schematics reading, and see if that inspires any further actions.
rick *eyes the Arduino on his desk*
I have had four or five cases similar to yours.All unresponsive in some way or another and all had to do with the 5V supply in some way. Voltage, ripple, bad filtering... That the problem gets worse within the first minute or so as things "warm up" is also in my notes.
And thanks Jacques.
Thanks Martin!
I now notice the Display circuit has a connection to RESET, and also ST-BY to the processors, but I don't understand yet what is happening there. I could replace C3 (small 4µ7 tantalum), and even C7 (Vss elco, 10µ) just as a trial. Those circuits are not fed by the 5V but by the 12V I think..
Things that take some time to fail (that warm up minute) smell like capacitor trouble to me.